Interpreting Microsoft's Big Data

"The large data platform installed in Windows Server and System Center is called Microsoft Hdinsight Server and is installed on Windows Azure by the Microsoft Hdinsight Service" This definition comes from an MSDN blog, which may seem abstract, TechEd 2012 Technical Conference site, Microsoft Asia-Pacific Research and Development Group chief technology Officer Sun Boke's speech, for everyone demo demo hdinsight application scenario.

Users of Excel can read Hadoop data in ODBC

"The first step in Microsoft's management of relational data, relational data and data flows is to build a platform under which all types of data can be integrated," Sun Boke in an interview with 51CTO reporters. The second step is to provide a tool for all data to be cleaned up and analyzed. We believe that all insights come from the relevance of data to data. "The current case in the Big data field, about the impact of US oil price volatility on car sales, is taking advantage of the market insights generated by the correlation between the two data."

Technology, Microsoft's important advantage is to help consumers use their most familiar tools to carry out applications. As you can see in the demo, either Hadoop on Windows Server or Hadoop on Windows Azure allows users to read the Hadoop platform's data through Excel. And in the Excel environment, the integration analysis of structured and unstructured data. Sun Boke says current applications can support tools such as Excel, PowerPivot for Excel, and power view.

In some of the information that users complete the installation of the Hive ODBC driver, you can see the new features in Excel Hive Query, by entering the Hadoop platform data source path to be analyzed, in an Excel environment, in ODBC mode, Read data from the Hadoop platform, and the results are stored in Excel or SQL Server in datasheet form or cube. Microsoft has repeatedly put forward the compatibility and attention to the Hadoop platform, this demonstration also becomes the focus in TechEd 2012, because Microsoft once again pushes the big data application directly to the user.

Deep collaboration with Apache Hadoop

Bing, Microsoft's search technology, originally had the concept of MapReduce distributed computing. However, Microsoft has also chosen to support more and more companies to start using Apache Hadoop, and as the core of unstructured data processing architecture. Enables organizations to work with unstructured data in the Windows environment for the Hadoop platform.

"We are primarily based on Windows Server and Windows Azure, while working well with Hortonworks platform," Sun Boke specifically to reporters: "Hadoop, Hdinsight technology, A more open approach is needed to advance with partners. Now, including PHP, MySQL, WordPress can be run on Windows Azure, more and more open source technology will also appear on the platform of Microsoft. In the on-site technical demo, Microsoft also demonstrated the streaming to iOS process specifically for mobile Services and media Services on Windows Azure.

As and integrated machine common layout large data

Memory computing and All-in-one is a hot spot for big data. At the pass annual summit hosted by the SQL Server user group in November this year, Microsoft Vice President Ted Kummert presented as computing as one of the core elements of Microsoft's data platform delivery strategy, Enables users to analyze various types of data while accelerating data access time. Ability to write data directly to RAM to eliminate performance bottlenecks.

In a Ted Kummert blog, Microsoft has been providing as technology in SQL Server since 2010, and the code for this project is "Hekaton", which will be provided as an important upgrade module for SQL Server. However, it is currently only used as a preview in a small number of users. "Hekaton" will improve Microsoft's In-memomory data portfolio, while spanning data analysis and business transaction scenarios. There will be a breakthrough performance improvement, and it's built into SQL Server, so companies don't need to buy extra hardware or software, and they can easily migrate existing applications, enabling these applications to achieve a breakthrough in performance.

"An online gaming website in Europe, when a game is staged, can be watched online by hundreds of millions of fans, who want to enter the trading platform with a series of real-time operations on the web." This is a big technical challenge for the Web site's providers, "Sun Boke, using a customer example to demonstrate a breakthrough in memory computing technology, and he said:" We have improved the efficiency of the entire online transaction 15~20 times with memory technology. Hekaton is a Latin name and is a hundredfold. We designed this technology at the very beginning, we hope to achieve a hundredfold efficiency, although it has not been achieved, but we will continue to upgrade as technology, hoping for a better breakthrough. ”

Ted Kummert also mentioned Microsoft Parallel Data Warehouse All-in-one PDW at the Pass annual summit. is for enterprise Data Warehouse, highly scalable design of soft and hard integrated equipment, the use of "large-scale parallel processing" (MPP) architecture. In data processing, parallel data warehouses based on SQL Server 2012 provide a new polybase data processing technology, xvelocity storage technology to meet the needs of real-time data warehouse, high-density Direct checkmark Storage, storage capacity increase 7 times times , the horizontal extension enables linear scaling from several terabytes to 6PB.

The technology for data has never been more fascinating than the fact that large data runs through applications, data centers, and clouds, ultimately into a capacity. This ability is to bring about change in life, work and thinking, and we interpret large data and interpret the world.

(Responsible editor: The good of the Legacy)